# HALO Uptime > Whether each part of HALO is up, and how often it has been down. Public JSON, no key needed. Names of creators are left > out unless you send an agent token (Authorization: Bearer agent..; ask HALO Uptime's owner for one). - [Everything, short](/status.json): {ok, overall: up | degraded | down | waiting, ranAt, groups: [{group, state, parts, down}]} - [Every part](/api/status): {overall, parts: [{id, group, label, state, since, text?, page, uptime: {d1, d7, d30, d90}, bars}], incidents} - [One part](/api/part/her-app): its state, uptime, 90 days of minutes up and down, and its last 50 incidents state is up, down or waiting (not seen yet). uptime figures are percentages, null before any minute was counted. bars is one character a day for 90 days, oldest first: space not watched, 0 up all day, 1 down under 15 min, 2 under an hour, 3 longer. The page checks every minute from Cloudflare, so it keeps answering when HALO's box is down: ranAt older than 3 minutes means the checker itself has stopped.